This study delves into the effects of blast loads on reinforced concrete columns, utilizing both Ductility-based and Residual Strength-based Evaluation Methods to assess their impact. The investigation specifically focuses on how the ratios of longitudinal and transverse reinforcement, as well as axial load ratios, influence the structural blast resistance. It was found that increased longitudinal reinforcement markedly enhances blast resistance, effectively reducing both displacement and strength damage. Similarly, a higher ratio of transverse reinforcement significantly boosts the lateral resistance of the columns. Conversely, elevated axial load ratios were observed to heighten displacement-based damage, attributable to the P-Δ effect, thus complicating the structural response to blast impacts. A comparative analysis between the two evaluation methods reveals a low consistency in their results, suggesting discrepancies that could impact the reliability of structural assessments under blast conditions. This inconsistency highlights the necessity for an integrated evaluation approach that combines both ductility and strength aspects to provide a more reliable and comprehensive assessment of blast resistance. Such an approach is essential for advancing current methodologies and ensuring the safety and integrity of structures subjected to explosive loads.

Acknowledgments: This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS-2024-00348713).

Numerous educational facilities remain at risk from seismic activity, having been erected prior to the institution of obligatory seismic design regulations. The focal point of this paper is the deployment of machine learning techniques to establish a swift and efficient reinforcement planning algorithm. This algorithm leverages minimalistic data inputs to ascertain the optimal fortification approach for school buildings from the 1980s that do not meet contemporary ductile design standards. By integrating a decision tree (DT) model, this investigation provides a conservative yet precise prediction of potential failure modes in concrete columns. The study further refines a strategy for determining optimal reinforcement tactics, taking into account the ductility enhancement through confinement ratio (CR) and the increase of stiffness via stiffness ratio (SR). The assessment of column failure types in relation to variations in CR and SR facilitates the formulation of a comprehensive retrofit plan for schools with masonry infill walls, delineating the upper limits of applicable SR and the permissible increment range based on CR thresholds. Compared to conventional analytical methods, the proposed retrofit planning technique significantly expedites the retrofit process, ensuring safer educational environments in seismically active zones. This work contributes to the field by providing a novel framework for seismic retrofitting that can be applied to vulnerable structures, as part of broader efforts to enhance seismic resilience in legacy building stock.

We introduce a novel approach, the Neural Network-Encoded Signed Distance Field (NetSDF), for the representation of shapes and computational particle mechanics in granular materials. Our method utilizes a neural network to learn and depict a Signed Distance Field (SDF), which defines a mapping from a point to a signed distance. Specifically, the neural network takes point coordinates and a latent code representing a single shape as inputs, generating the signed distance from the point to the particle surface. The sign distinguishes between the interior and exterior of a particle, making the zeroth isosurface of the SDF an accurate representation of the particle surface. Upon training the NetSDF with a designated set of particle samples, it can effectively represent an entire class of particles exhibiting the characteristic morphology of the granular material. Our results demonstrate the NetSDF's proficiency in accurately representing irregular-shaped particles and generating virtual particles within the same class. Moreover, the NetSDF seamlessly integrates with the SDF-based discrete element method (Lai et al., 2022, Computational Mechanics), showcasing notable advantages in terms of memory consumption and computational efficiency.

Reference:

Lai, Z., Zhao, S., Zhao, J., & Huang, L. (2022). Signed distance field framework for unified DEM modeling of granular media with arbitrary particle shapes. Computational Mechanics, 70(4), 763-783.

Seismic damages of nonstructural components such as partition walls and ceilings installed in buildings can be monitored to increase occupant safety and to perform immediate repairs. Since damaged locations of the nonstructural components are difficult to predict compared to structural components, traditional vision-based displacement sensors (VDSs) with markers limited to single-point measurement are not suitable for detecting the seismic damages. This study aimed to develop a shaped-based VDS measuring multi-points displacement optimized for seismic damage detections of the nonstructural components. Using shape information (i.e., convex hull) of the nonstructural components in region of interests to extract and track multiple feature points at user-desired locations, the seismic damage of the nonstructural components can be detected and monitored with a single camera. From the shake table test on a two-story moment frame installed with partition walls and ceiling, the applicability of the proposed vision system is investigated and discussed. The result demonstrates that the proposed vision system can detect the damage location of the nonstructural components without any additional equipment.